1 / 36

Levon Stepanian, Angela Demke Brown Computer Systems Group

Inlining Java Native Calls at Runtime (CASCON 2005 – 4 th Workshop on Compiler Driven Performance). Levon Stepanian, Angela Demke Brown Computer Systems Group Department of Computer Science, University of Toronto Allan Kielstra, Gita Koblents, Kevin Stoodley IBM Toronto Software Lab.

wyatt
Download Presentation

Levon Stepanian, Angela Demke Brown Computer Systems Group

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Inlining Java Native Calls at Runtime(CASCON 2005 – 4th Workshop on Compiler Driven Performance) Levon Stepanian, Angela Demke Brown Computer Systems Group Department of Computer Science, University of Toronto Allan Kielstra, Gita Koblents, Kevin Stoodley IBM Toronto Software Lab

  2. In a nutshell • Runtime native function inlining into Java • Optimizing transformations on inlined JNI calls • Opaque and binary-compatible while boosting performance Java Code Native Function Call Native Code

  3. In a nutshell • Runtime native function inlining into Java • Optimizing transformations on inlined JNI calls • Opaque and binary-compatible while boosting performance Java Code Native Code inlined

  4. In a nutshell • Runtime native function inlining into Java • Optimizing transformations on inlined JNI calls • Opaque and binary-compatible while boosting performance Java Code Native Code inlined + optimized

  5. Motivation • The JNI • Java’s interoperability API • Callouts and callbacks • Opaque • Binary-compatible Java App JVM + JIT JNI Callout Callback Native (Host) Environment Native App/Lib

  6. Motivation • The JNI • Pervasive • Legacy codes • Performance-critical, architecture-dependent • Features unavailable in Java (files, sockets etc.)

  7. Motivation • Callouts run to 2 to 3x slower than Java calls • Callback overheads are an order of magnitude larger • JVM handshaking requirements for threads leaving and re-entering JVM context • i.e. stack switching, reference collection, exception handling • JIT compiler can’t predict side-effects of native function call

  8. Our Solution • JIT compiler based optimization that inlines native code into Java • JIT compiler transforms inlined JNI function calls to constants, cheaper operations • Inlined code exposed to JIT compiler optimizations

  9. Infrastructure • IBM TR JIT Compiler + IBM J9 VM • Native IL to JIT IL conversion mechanism • Exploit Native IL stored in native libraries • W-Code to TR-IL at runtime Machine code + TR JIT Static compiler IL

  10. Outline • Background Information  • Method • Results • Future Work

  11. Sample Java Class class SetFieldXToFive{ public int x; public native foo(); static{ System.loadLibrary(…); } }

  12. Sample Java Class class SetFieldXToFive{ public int x; public native foo(); static{ System.loadLibrary(…); } }

  13. Sample Native Code JNIEXPORT void JNICALL Java_SetFieldXToFive_foo (JNIEnv * env, jobject obj){ jclass cls = (*env)->GetObjectClass(env,obj); jfieldID fid = (*env)->GetFieldID(env,cls,“x","I"); if (fid == NULL) return; (*env)->SetIntField(env,obj,fid,5); } GOAL: obj.x = 5

  14. Sample Native Code JNIEXPORT void JNICALL Java_SetFieldXToFive_foo (JNIEnv * env, jobject obj){ jclass cls = (*env)->GetObjectClass(env,obj); jfieldID fid = (*env)->GetFieldID(env,cls,“x","I"); if (fid == NULL) return; (*env)->SetIntField(env,obj,fid,5); } GOAL: obj.x = 5 SetFieldXToFive

  15. Sample Native Code JNIEXPORT void JNICALL Java_SetFieldXToFive_foo (JNIEnv * env, jobject obj){ jclass cls = (*env)->GetObjectClass(env,obj); jfieldID fid = (*env)->GetFieldID(env,cls,“x","I"); if (fid == NULL) return; (*env)->SetIntField(env,obj,fid,5); } GOAL: obj.x = 5

  16. Sample Native Code JNIEXPORT void JNICALL Java_SetFieldXToFive_foo (JNIEnv * env, jobject obj){ jclass cls = (*env)->GetObjectClass(env,obj); jfieldID fid = (*env)->GetFieldID(env,cls,“x","I"); if (fid == NULL) return; (*env)->SetIntField(env,obj,fid,5); } GOAL: obj.x = 5

  17. Sample Native Code JNIEXPORT void JNICALL Java_SetFieldXToFive_foo (JNIEnv * env, jobject obj){ jclass cls = (*env)->GetObjectClass(env,obj); jfieldID fid = (*env)->GetFieldID(env,cls,“x","I"); if (fid == NULL) return; (*env)->SetIntField(env,obj,fid,5); } GOAL: obj.x = 5

  18. Native Inlining Overview • Inliner detects a native callsite • Extracts and converts Native IL to JIT IL • Identifies inlined JNI calls • Transforms inlined JNI calls • Finishes inlining

  19. Method – Step 1 TR JIT 1. Inliner detects a native callsite Inliner Java Code foo(){…} (Native code) Call to obj.foo()

  20. Method – Step 2 • Inliner detects a native callsite • Extracts and converts Native IL to JIT IL Native IL JIT IL

  21. Method – Step 3 • Inliner detects a native callsite • Extracts and converts Native IL to JIT IL • Identifies inlined JNI calls JIT IL /* call to GetObjectClass */ … /* call to GetFieldID */ … /* call to SetFieldID */ … Pre-constructed IL shapes

  22. Method – Step 4 jclass cls = (*env)->GetObjectClass(env,obj); jfieldID fid = (*env)->GetFieldID(env,cls,“x","I"); if (fid == NULL) return; (*env)->SetIntField(env,obj,fid,5); • Inliner detects a native callsite • Extracts and converts Native IL to JIT IL • Identifies inlined JNI calls • Transforms inlined JNI calls

  23. Method – Step 4 jclass cls = (*env)->GetObjectClass(env,obj); Constant: SetFieldXToFive class data structure • Inliner detects a native callsite • Extracts and converts Native IL to JIT IL • Identifies inlined JNI calls • Transforms inlined JNI calls jfieldID fid = (*env)->GetFieldID(env,cls,“x","I"); if (fid == NULL) return; Constant: Offset of field “x” (*env)->SetIntField(env,obj,fid,5); JIT IL: obj.x = 5

  24. The Big Picture TR JIT Before Native Inlining & Callback Transformations • Inliner detects a native callsite • Extracts and converts Native IL to JIT IL • Identifies inlined JNI calls • Transforms inlined JNI calls • Finishes inlining Inliner Java Code foo(){…} (Native code) Call to obj.foo()

  25. The Big Picture TR JIT After Native Inlining & Callback Transformations • Inliner detects a native callsite • Extracts and converts Native IL to JIT IL • Identifies inlined JNI calls • Transforms inlined JNI calls • Finishes inlining Inliner Java Code foo(){…} (Native code) obj.x = 5

  26. The Big Picture TR JIT After Native Inlining & Callback Transformations • Inliner detects a native callsite • Extracts and converts Native IL to JIT IL • Identifies inlined JNI calls • Transforms inlined JNI calls • Finishes inlining Inliner Java Code obj.x = 5

  27. Outline • Background Information  • Method  • Results • Future Work

  28. Experimental Setup • Native function microbenchmarks • Average of 300 million runs • 1.4 GHz Power4 setup • Prototype implementation

  29. Cost of IL Conversion • 5.3 microseconds per W-Code

  30. Inlining Null Callouts • Null native method microbenchmarks • Varying numbers of args (0, 1, 3, 5) • Complete removal of call/return overhead • Gain back 2 to 3x slowdown • confirmed our expectations

  31. Inlining Non-Null Callouts • smaller speedups for natives performing work • instance vs. static speedup

  32. Inlining & Transforming Callbacks • Reclaim order of magnitude overhead

  33. Data-Copy Speedups Speedup (X) Array Length • Transformed GetIntArrayRegion

  34. Exposing Inlined Code To JIT Optimizations FindClass GetMethodID NewCharArray GetArrayLength

  35. Conclusion • Runtime native function inlining into Java code • Optimizing transformations on inlined Java Native Interface (JNI) calls • JIT optimize inlined native code • Opaque and binary-compatible while boosting performance • Future Work • Engineering issues • Heuristics • Larger interoperability framework

  36. Fin

More Related